I tried the Windows creator USB 2.0 Audio drivers, and they reported correct sample rates, but when playing a tune they only play for about 5 seconds then stop. So it seems this is a driver issue.
Btw, do you have any idea why the windows drivers don’t work?
@Magnus, huh i had same issues thought cause my laptop was connected via Wifi … all i was checking if it clicks on sample rate change. but I agree - it seems to have a problem with ‘native’ windows 10 usb 2.0 drivers. i did’t try anything else as i did not try drivers that came wiht a CD … actually i dont have a CD to read it from 
Hi Magnus - no popping for me. Enabled DSP, custom sample rate and switched between 44.1 and 48 repeatedly. Can hear the music change subtly (volume mainly) but no pops.
Something I did have though was when I first enabled custom sample rate (without making any changes) I got a repeated noise and then Roon lost the audio zone. I had to unplug it and re-enable it.
Might be a difference between Mac and Windows/Linux then, or we have different batches.
I tried to up-sample to DSD265 right now, sound a little different (sort of softer), but the CPU demand was around 10% so I have now settled on 384khz which demands 1.5 - 2% CPU. Sounds pretty awesome, and no clicks.
They have a bug where the driver fails to report that the clock is advancing during playback. So we load up the first few seconds, but use their clock to determine how fast to send the remaining audio. If the clock gets stuck at zero, as with this driver, then the remaining audio never makes it to the driver. Basics.ly, we think that the driver is taking forever to start playback.
We couldn’t find a way to work around this bug from our side without breaking synchronized playback. Since in practice other drivers are always available, we can wait before trying to do anything too desperate.
Most apps don’t care, and ignore the clock. RAAT does a lot more with clock management and so we need this stuff to be working. I assume that at some point Microsoft will figure this out and fix it.
Cheers for the answer, but I would not hold my breath for Microsoft to fix it. Look at how long it took before they even tried to support USB Audio 2.0 
My guess is that they supported it for because it isn’t practical to have every audio device manufacturer go back and make ARM drivers for their ARM based products, but they still want standard devices to work. Before then there wasn’t really a business reason for them to do it since manufacturers were already accustomed to making drivers.
In my experience, bugs like this tend not to survive more than one Windows release. There are just so many people using Windows that any legit bug turns into a noticeable volume of reports, and most crucially, reports from large customers with expensive support contracts—I.e. not us.
ok i tried Mac as well. If i don’t use exlusive mode and go via CoreAudio then there are no pops/clicks but explanation is simple - it’s all downsmapled to same bitrade by mac audio system. no surprise. MQA doesn’t work (i mean you do get sound/music bit it’s not recognised as MQA cause Mac subsystem does not make it lossless and/or it re-samples).
If i choose Exclusive mode (and then i see all the options e.g. DSD64/128/256 via DOP PCM thorugh 768kHz) i do get pops/clicks.
@anon73739233 i doubt you have been running non-exclusive mode on mac ? otherwise given our serial numbers are so close it sounds like HW is the same batch though they behave differently.
I did get iFi iPower so that it always has clean power rather then pulling it from player (mac or sbc) can’t say i heard huge change in sound but pops are bit less frequent … or i am imagening. Though if i want to get them all it takes on 44.1 and one 96khz song and flip between 2 every 0.5-1sec.
Not that i care a lot about MQA right now - and upsampling either via roon or via squeezelite to 192 or 352 or 384 or even 705 and 768 sorts that out and i dont think my DSD material sounds worse … so the only thing that does not work in that scenario is MQA. So short term all of the complaignts on pops/clicks are real but easily avoidable. On top of that EVEN if it never clicked/popped i do like ReplayGain as I have mixed playlists with Tidal and my own material … so if i put MQA in there and have digital volume - MQA gets lost. If Meridian did not keep design closed i am 100% that even that could be fixed in SW … but it’s easier to tell dumbasses that HW decoding is a must (e.g. FPGAs at DACs can do the job that computer can not … and all FPGA’s do to bits and bytes is same algorithm that would run even much faster on most todays HW including SBCs). anyhow, that is it from my side !
Thanks for the update. I’m running mine in exclusive mode.
Tried some real DSD and it really is an improvement over PCM, so silky smooth. I am also a little surprised that I can so easily hear the improvement, since I run a pair of powered speakers with DPS in them (i.e. they do A/D and D/A internally). I thought that it would nullify any higher sample rate improvement but it does not, at least not totally.
Listening on Dark side of the moon now (and if you don’t know what group it it, shame on you 
) in DSD 64, sounds awesome!
Even up-sampling PCM to 384 is a nice improvement over 44.1 (44.1 sound hard and edgy in comparison).
Despite the problems with pops on sample rate switch (which will be fixed anyway), I am totally happy with my little DAC (btw, it looks a little like a toy-DAC, something audiophiles might give to their 5 year old kid to play with )
Tried some real DSD and it really is an improvement over PCM, so silky smooth. I am also a little surprised that I can so easily hear the improvement, since I run a pair of powered speakers with DPS in them (i.e. they do A/D and D/A internally). I thought that it would nullify any higher sample rate improvement but it does not, at least not totally.
I have had the same concerns after running DSD on my Marantz dac into my Dynaudio active which converts the signal into digital before dsp and amplification. It sounds great, but it would be fun to know more about the possible technical disadvantages of multiple conversions between d/a a/d d/a. Anyone know any good resources where one could read some more?
Anyone know why I can’t up-sample in Roon to 768khz despite that the dac supports it and Roon reports that it supports 768? Probably time to bother @brian as always
I don’t see any reason why you shouldn’t be able to. I’ve seen it work before. More details?
I use the drivers that came with the DAC (UNI Project Driver v1.61.4), and everything except DSD512 is green when I use the “Work around drivers that misreport devices capabilittes” in device setup. But when I set Max PCM Rate in DSP - Sample Rate Conversion it only up-samples to 384. And if I select Custom I can select up to 384 khz and up to DSD256.
Funny thing: I can select 768khz as source sampling in Custom, but not select to sample up to 768khz.
Here is screenshot of the Custom sample conversion, with 44.1khz menu expanded:
No, no luck. But I got a good computer so I up-sample everything to DSD128 instead, takes about 4% CPU which means I can work or play on computer at the same time (processing speed at 7 in signal path). And it sounds better than PCM
@Dirk-Pitt - what are you plugging your S2 in to? I see it’s a ropieee endpoint (and using ALSA) wonder if that is a factor - I’m having the same experience as @Magnus
Magus - upsampling is ok. But if you play MQA now I bet it’s not MQA. With the S2 you almost want a DSP on/off switch on the player in Roon so you can flip between MQA and not as tracks change.
Thanks. I’ll try plugging mine into the USB of a Pi and see how that goes - if it makes a difference. Today I plugged the S2 into 2 laptops. One won’t let me upsample to DSD, the other will.
My ROCK box ( not a NUC ) has not enough power for that 
